Lung cancer detection system..
in [Matlab]
|
Prev: probability density function vs likelihood
Next: regarding Sensorless Control of Permanent Magnet Synchronous Motor Based on Extended Kalman Filterl
From: Gaya on 22 May 2010 07:03 Dear all, I'm currently developing a cancer detection system. I use digital chest xray images as the input to the system. I extract 6 first order statistic texture features and 4 second order statistic features. Is that enought to consider these features as the input to train the neural network..?
From: Mark Shore on 22 May 2010 08:05 "Gaya " <gayaudeshani(a)yahoo.com> wrote in message <ht8dl7$bpn$1(a)fred.mathworks.com>... > Dear all, > > I'm currently developing a cancer detection system. > I use digital chest xray images as the input to the system. I extract 6 first order statistic texture features and 4 second order statistic features. > > Is that enought to consider these features as the input to train the neural network..? This is not a MATLAB question.
From: John D'Errico on 22 May 2010 08:18 "Gaya " <gayaudeshani(a)yahoo.com> wrote in message <ht8dl7$bpn$1(a)fred.mathworks.com>... > Dear all, > > I'm currently developing a cancer detection system. > I use digital chest xray images as the input to the system. I extract 6 first order statistic texture features and 4 second order statistic features. > > Is that enought to consider these features as the input to train the neural network..? As Mark points out, this is not a matlab question. Nor can anybody answer (on any newsgroup or forum) if you will ever have ENOUGH data. The answer is that "enough" can only be determined by you, knowing your goals. The more data that you provide, the better in general will be your results. John
From: Gaya on 22 May 2010 10:17 "John D'Errico" <woodchips(a)rochester.rr.com> wrote in message <ht8i1s$l50$1(a)fred.mathworks.com>... > "Gaya " <gayaudeshani(a)yahoo.com> wrote in message <ht8dl7$bpn$1(a)fred.mathworks.com>... > > Dear all, > > > > I'm currently developing a cancer detection system. > > I use digital chest xray images as the input to the system. I extract 6 first order statistic texture features and 4 second order statistic features. > > > > Is that enought to consider these features as the input to train the neural network..? > > As Mark points out, this is not a matlab question. Nor can > anybody answer (on any newsgroup or forum) if you will > ever have ENOUGH data. The answer is that "enough" can > only be determined by you, knowing your goals. The more > data that you provide, the better in general will be your > results. > > John Thank you but sometimes people can answer by their experience.. And sorry I am new to this newsgroup. I use Mathlab to develop this thatz why I asked that question...
From: John D'Errico on 22 May 2010 10:29
"Gaya " <gayaudeshani(a)yahoo.com> wrote in message <ht8p10$d3d$1(a)fred.mathworks.com>... > "John D'Errico" <woodchips(a)rochester.rr.com> wrote in message <ht8i1s$l50$1(a)fred.mathworks.com>... > > "Gaya " <gayaudeshani(a)yahoo.com> wrote in message <ht8dl7$bpn$1(a)fred.mathworks.com>... > > > Dear all, > > > > > > I'm currently developing a cancer detection system. > > > I use digital chest xray images as the input to the system. I extract 6 first order statistic texture features and 4 second order statistic features. > > > > > > Is that enought to consider these features as the input to train the neural network..? > > > > As Mark points out, this is not a matlab question. Nor can > > anybody answer (on any newsgroup or forum) if you will > > ever have ENOUGH data. The answer is that "enough" can > > only be determined by you, knowing your goals. The more > > data that you provide, the better in general will be your > > results. > > > > John > > Thank you > but sometimes people can answer by their experience.. > And sorry I am new to this newsgroup. I use Mathlab to develop this thatz why I asked that question... Again, our experience is not even relevant. We cannot say if you have enough data. You will never know that you have enough data. The more data you have, the better your estimates. John |